PROJECT SUMMARY/ABSTRACT To address more than a quadrupling of death rates from opioid overdose between 2000 and 2015, federal and state agencies have promoted clinician education and guidelines to reduce risky prescribing. Previous research has identified prescription patterns associated with elevated risk of opioid overdose, yet the relationship between patient and environmental factors, prescription use/misuse trajectories, and overdose likelihood remains largely unknown. This proposal, submitted in response to PAR-16-234 (Accelerating the Pace of Drug Abuse Research Using Existing Data), will develop comprehensive models for assessing opioid overdose risk, filling critical gaps in understanding of how prescription opioid use/misuse changes over time, how such changes affect overdose risk, which patients are most vulnerable to risky patterns, and what role household- and community-level prescription risk plays in overdose. The specific aims are: 1. Model effects of patient demographic and clinical characteristics and patient prescription patterns and their interactions on opioid-involved overdose (fatal or nonfatal). 2. Determine the effect of household-level prescription availability on opioid overdose. 3. Determine the effect of community-level prescription availability on opioid overdose. A key strength of our study is our novel linked dataset: the Oregon Comprehensive Opioid Risk Registry (CORR), which links prescription and clinical history across payers with diverse sources of overdose data, including data from the Oregon Prescription Drug Monitoring Program, Medicaid Claims, Vital Records, and Hospital Discharge registry, as well as All Payer/All Claims and Emergency Medical Services data. Our study will determine the odds of opioid-related overdose based on interactions of patient demographics, diagnoses/comorbidities, initial opioid prescriptions, household prescription risk levels, and community prescription risk levels. The study will use models to examine how risk builds over time and identify prescription patterns that portend increased risk at an early stage. Innovation: Our study creates a novel linked dataset and applies a complex analytic approach to radically expand understanding of patients' individual risk environments. Significance: This study will inform clinical practice by generating new knowledge that can help identify the most at-risk patients and modify opioid prescribing decisions regarding them. Impact: Hierarchical models which combine individuals' prescription trajectories and clinical histories with household-level and community-level risk factors can be extended to other complex diseases in which the adverse outcomes occur as a result of effects acting at different levels.